Escolar Documentos
Profissional Documentos
Cultura Documentos
Chapter 2 Reading
instructions
E
Bias: Systemtic deviation from the true value
Design, Conduct, Analysis, Evaluation
-10
-7
Drog X - Placebo
-4
n=40
mm Hg
-10
Population mean
bias
= population mean
Drog X - Placebo
-7
-4
n=200
mm Hg
-10
-7
-4
N=2000
mm Hg
Detection bias
Analysis bias
Interpretation
bias
Temporal:
Drug X
Baseline
8 weeks
Y X
Variation
= Explaine + Unexplaine
in
d
d variation
observation
variation
s
Drug A
Drug B
Outcome
Y=B+x
B
Y=A+x
A
x=age
Confounding
Predictors
of
treatment
Confounder
s
Predictors
of outcome
A
Treatment
Treatment
allocation
allocation
Outcome
B
Example
Smoking Cigarettes is not so bad but watch
out for Cigars or Pipes (at least in Canada)
Variable
20.5
Cigar or pipe
smokers
35.5
Cochran, Biometrics 1968
Example
Cigar or pipe
smokers
20.5
35.5
Average age
50.5
65.9
54.9
Example
Cigar or pipe
smokers
20.5
35.5
Average age
54.9
50.5
65.9
Adjusted
20.2
mortality rate*
26.4
24.0
Confounding
The effect of two or more factors can not be separated
Example: Compare survival for
surgery and drug
Survival
Looks ok
Surgery at time 0
Surgery only if healty enough
but:Patients in the surgery arm may take
drug
Complience in the drug arm May be
Time
Confounding
Can be sometimes be handled in the design
Example: Different effects in males and females
Imbalance between genders affects result
Stratify by gender
A
Gender
R
B
Balance on average
B
A
Always balance
Interaction
The outcome on one variable
depends on the value of another
variable.
Example Interaction between two drugs
A
B
A=AZD1234
Wash
out
R
B
B=AZD1234 +
A
Clarithromycin
Interaction
Example: Drug interaction
Plasma concentration (mol/L)
linear scale
Mean
5
AZD1234
AZD0865 alone
Combination
of clarithromycin
AZD1234
and AZD0865
2
1
0
0
12
16
20
24
AUC AZD1234:
19.75 (mol*h/L)
Interaction
Example: Treatment by center interaction
Treatment difference in diastolic blood pressure
15
10
5
mmHg
0
-5
10
15
20
25
30
-10
-15
-20
-25
Ordered center number
Descriptive statistics
Inferential statistics
Explorative statistics
Confirmatory analysis:
Hypothesis
Results
Conclusion
Hypothesis
Conclusion?
Estimate
p-value
Confidence interval
Results
Interpretation
X
,
R
Statistical model: Observations 1
n
P :
Reject H0 if X S c R n
P X S c | 0
Rejection region
Significance level
Confidence intervals
Let X, *
1 if H 0 : * rejected
0 if H 0 : not rejected
*
(critical function)
Confidence set:C X : X, 0
The set of parameter values correponding to
hypotheses that can not be rejected.
C X
A confidence set is a random subset
covering the true parameter value with
1 at least
probability
.
Example
Objective: To compare sitting diastolic blood pressure (DBP) lowering effect
of hypersartan 16 mg with that of hypersartan 8 mg
4
Parameter space:
R
1 2 3 0
R
Null hypoteses (subsets of
H01: 1 = 2 (DBP)
H02: 1 = 2 (SBP)
H03: 2 = 3 (DBP)
H04: 2 = 3 (SBP)
):
Example contined
Hypothesis
Variable
1: 16 mg vs 8 mg
Sitting
DBP
2: 16 mg vs 8 mg
3: 8 mg vs 4 mg
4 : 8 mg vs 4 mg
This is a t-test
CI (95%)
p-value
-3.7 mmHg
[-4.6, -2.8]
<0.001
Sitting
SBP
-7.6 mmHg
[-9.2, -6.1]
<0.001
Sitting
DBP
-0.9 mmHg
[-1.8, 0.0]
0.055
-2.1 mmHg
[-3.6, -0.6]
0.005
Sitting
SBP the
where
Rejection region:X : T
-c
LS Mean
T X
-4.6
-2.8
1 2
Estimation of effect
p-value
10
1.94 mmHg
0.376
100
-0.65 mmHg
0.378
1000
0.33 mmHg
0.129
10000
0.28 mmHg
<0.0001
100000
0.30 mmHg
<0.0001
Drug
Healing rate
gastroprazole 40
mg
87.6%
mygloprazole 30
mg
84.2%